Summary

  • The public record supports a precise identity boundary. Northern Capital Gateway is the Pulkovo Airport operator under a long public-private partnership, with a large transport asset, public airport tariffs, non-aeronautical tenant services and a small autonomous system. It does not support treating the company as a conventional regional ISP. AS198626 matters because it proves direct routing control over a 2,048-address IPv4 block and visible upstream relationships, not because it proves a wholesale access customer base.
  • The first paid interconnection or access port would need to be measured as retained contribution, not gross resale revenue. Pulkovo's airport economics create potential internal demand and tenant aggregation, but published BGP and PeeringDB evidence shows no downstreams, no public exchange presence and only a narrow supplier set. The base judgment is therefore conditional and conservative: the airport can justify resilient in-house connectivity and perhaps limited campus access services, but a standalone gateway business only works if utilisation is high, customers are captive enough to pay for service assurance and management can avoid overbuilding against direct carrier alternatives.

Start with the first paid port

The right way to read Northern Capital Gateway is to begin with one paid port. Not the airport terminal, not the concession, not the 20.9 million passenger headline, and not the comfort of a high credit rating. One port forces the issue. A tenant, airline office, service partner, cargo-related user, advertising system, hotel or operational counterparty pays the airport operator for connectivity or an access interface. The invoice may include access, local cabling, managed Wi-Fi, operational support, security segregation, billing administration, service desk handling and perhaps an internet handoff.

From the outside, the gross line may look like telecom revenue. Economically it is not all telecom contribution.

The carrier portion is pass-through. If the operator is buying upstream capacity from ER-Telecom, MegaFon or another direct carrier, the external bandwidth cost belongs above the contribution line. If the tenant can order business internet directly from MTS, Rostelecom, MegaFon or another local supplier, the airport cannot price the service as if it owned a protected monopoly over IP transit.

The airport can charge for what the direct carrier cannot conveniently deliver: inside-airport access, controlled installation, security coordination, operating-hours support, faster fault handling inside a restricted environment, accountability to airport operating procedures and the removal of friction for counterparties that need to function at Pulkovo. The question is whether that retained layer is wide enough to pay for ports, upstream capacity, equipment, staff, monitoring, replacement cycles and the risk of service failure in a critical transport venue.

That framing is deliberately stern because the company is not a clean regional ISP case. Northern Capital Gateway's public pages describe an airport operator. The Pulkovo operator page says the company has managed the airport since 29 April 2010, implements Russia's first aviation public-private partnership without budget funds, and invested about EUR 1.2 billion from 2010 through 2015 in the airport's development.

The public-private partnership page gives the legal shape: the city and the city-owned Pulkovo Airport company transferred the airport property complex into long-term lease for 30 years; Northern Capital Gateway receives the exclusive right to provide airport services on airport territory; and the operator must manage, provide airport services, invest and pay a concession payment with fixed and variable parts. That is an airport-services contract, not a telecom concession.

Still, the network evidence is real. RIPE's member list identifies Limited Liability Company "Northern Capital Gateway" in Russia. RIPE Stat shows AS198626 as announced and held as NCG-AS by the same company. The announced-prefixes data shows 37.205.16.0/21, or 2,048 IPv4 addresses, visible in the recent observation window. RIPE routing-status data shows one IPv4 prefix, no IPv6 announced space, visibility across all queried IPv4 RIS peers at the time checked, and two observed neighbours.

The RIPE whois record names AS20632 and AS9049 as import sources and export targets, which corresponds with MegaFon and ER-Telecom evidence in third-party network pages. PeeringDB's API, however, returns no network entity for ASN 198626. That absence is not proof of no private commercial services, but it is a strong sign that the company is not presenting itself to the peering market as a normal open interconnection participant.

The economic question therefore has two layers. The first is internal: does owning the autonomous system and address space make Pulkovo's operational estate more resilient, controllable and economical than buying all service as a retail customer? The answer is likely yes, within limits, because a large airport is not an ordinary office. It has airline counters, passenger processing, retail systems, access control, operations centers, public information displays, tenant environments, airside and landside coordination, security procedures and high peak loads.

The second layer is external: can that network position support meaningful paid access or interconnection revenue? The answer is unproven. It can be plausible for a campus utility. It is not yet visible as a wholesale gateway.

The operating asset is valuable, but it is not the same as a carrier moat

Pulkovo gives Northern Capital Gateway a rare operating surface. The official airport page describes one of Russia's large air transport nodes, around 4,000 employees, 1,350 hectares of airport area, two runways, 88 check-in desks, 107 passport-control booths with daily capacity stated at 57,000 people, 16 jet bridges, 110 aircraft stands, baggage belts, lifts, escalators and boarding gates.

The airfield page adds that the airfield is in the state register of civil aerodromes, has a certificate of state registration and suitability for operation, is fit for international flights, has two independent parallel paved runways, a taxiway system and fire protection category eight. These facts matter because connectivity at an airport is not a discretionary convenience. It sits beside physical safety, passenger throughput, regulated access and time-sensitive commercial operations.

The airport also has traffic. Pulkovo's own 2024 year-end release reported 20.9 million passengers, 16.7 million domestic and 4.2 million international, a 3 percent increase over 2023 and the airport's second consecutive year above the 2019 benchmark. It said 55 airlines operated 111 destinations during 2024, with Rossiya, Pobeda, Smartavia, Aeroflot and Nordwind leading Russian-carrier volumes and Turkish Airlines, Uzbekistan Airways, Belavia, Pegasus and Emirates among foreign-carrier leaders. It also reported more than 2.2 million passengers carried under the Moscow shuttle program in 2024.

The airport's airline-partner page is more promotional and says the catchment area exceeds 10 million people and the route network includes more than 160 Russian and international destinations operated by more than 70 airlines. The difference between the operational release and the general partner page is a useful warning: for unit economics, use current measured throughput and disclosed carrier concentration, not the broadest marketing number.

Large traffic does not automatically become data traffic revenue. Passenger volume can lift Wi-Fi use, digital advertising impressions, parking systems, biometric projects, passenger-flow analytics and tenant demand, but only a slice of that can be monetized as access. Airlines and large tenants may have their own national telecom procurement. Banks, payment processors, retailers, restaurants and logistics firms often prefer direct connections or managed services from their existing carriers.

A foreign or domestic airline may value one airport-managed circuit if it gets a counter open quickly, but it may not want to rely on the airport operator as its long-term primary network provider. The operator's leverage is convenience and controlled space, not a broad metro-fiber monopoly.

That is why the first paid port must be stripped down. Suppose a customer pays for a managed connection in the terminal or business center. The gross price must cover external capacity, cross-connect or access build, router and switch depreciation, security separation, help desk work, fault escalation, billing and a risk premium for airport-hours support. If the customer merely reimburses the carrier cost plus a small administration fee, the business is operationally useful but not a gateway profit engine.

If the airport can aggregate many small counterparties over shared infrastructure and sell a standardized, high-assurance product, then the retained margin can become interesting. The public evidence does not prove either outcome, so the only defensible judgment is to make utilisation and contribution the gating variables.

Route diversity is thin for a wholesale claim

The routing data is the most disciplined evidence because it is less promotional than corporate copy. AS198626 is announced. The active prefix is 37.205.16.0/21. RIPE Stat saw that route in early August 2026, with first-seen history dating back to May 2012 and with IPv4 visibility across the queried RIS peer set. The BGP footprint is therefore not a dead artifact. It is live enough to matter in an operational review.

But the footprint is small. One /21 gives 2,048 IPv4 addresses. That is ample for an airport campus, corporate services, public-facing systems and segmentation, especially when combined with private addressing and managed internal networks. It is not a sign of a regional access network serving large numbers of broadband customers. There is no visible IPv6 announcement in the RIPE routing-status data. For a public internet access business in 2026, the absence of IPv6 is not fatal, but it lowers the quality signal.

It suggests either a legacy enterprise network posture or a narrow operational network whose customers are not pushing the operator to present itself as a modern access ISP.

The neighbour set is also narrow. The RIPE whois record imports any routes from AS20632 and AS9049 and exports AS198626 to both. IPinfo summarizes one upstream and two peers, while RIPE's neighbour data observed two neighbours. IPIP's AS page likewise names AS198626 as NCG-AS and shows the same single IPv4 prefix and no IPv6 prefix. This is consistent with basic multi-homing or carrier redundancy. It is not consistent with a broad wholesale gateway that offers rich path choice, multiple transit providers, Internet exchange presence and a downstream customer cone. PeeringDB's empty response for ASN 198626 reinforces the point.

PeeringDB is not a complete census of private interconnection, but a serious interconnection seller usually has reason to make its presence, facilities, policy and contact surface findable there.

The cost implication is hard. A single airport-owned AS can be the right architecture for control and resilience. It can keep public addresses under the airport operator's policy, allow better inbound routing control, simplify certain security or remote-access designs, and reduce dependence on one retail ISP account. But every step toward a commercial gateway raises fixed costs. More upstreams mean more ports, more contracts, more monitoring, more routing skill and more emergency procedures. Exchange presence means transport to the exchange, switch ports, memberships, cross-connects and operations.

Customer access means service-level commitments, demarcation, installation rules and dispute handling. In a normal carrier business, these costs are diluted over many access customers or high traffic volume. In an airport estate, they are diluted only if many tenants and operational partners buy the service at prices above pass-through.

The current evidence points to an enterprise or campus network first, and a commercial access platform only as a possible adjacency. That does not make the network irrelevant. On the contrary, it makes the first port financially revealing. If one paid port has high installation friction, heavy support needs and low retained contribution, adding customers makes complexity worse. If the first port can be provisioned from existing internal distribution, uses a standard offer, rides already-purchased resilient upstream capacity and creates meaningful margin, then the airport's physical control can become an economic asset.

The airport concession absorbs capital before the network can become strategic

Northern Capital Gateway's capital story is dominated by airport infrastructure. IFC's project disclosure described a project company selected for a 30-year public-private partnership to expand, develop, operate and maintain Pulkovo, then Russia's fourth-largest airport by 2009 traffic, with a new terminal and airside and landside investments targeting 17 million passengers per year. NIB's loan page records a EUR 50 million, 15-year loan for the reconstruction and expansion project and estimates total project costs around EUR 1.19 billion, with about 60 percent financed through loans.

EBRD's project page records a EUR 100 million A-loan and project cost of EUR 1.2 billion, with the project company building, expanding, financing and operating airport facilities, except cargo terminal and business aviation, under the PPP. Eurasian Development Bank recorded a USD 90 million loan. The old project-finance story is not a telecom story, but it tells us where management's capital discipline was formed: large fixed assets, long concession life, debt service, passenger-flow risk and regulated infrastructure obligations.

The present credit story remains airport-led. Pulkovo's 2025 rating announcement said ACRA upgraded the company to AA+ with stable outlook and Expert RA confirmed AA+ with stable outlook. The same announcement highlighted very low leverage, high interest coverage, strong cash flow, high profitability, high liquidity and low regulatory risk, with 2024 passenger traffic of 20.9 million and first-half 2025 passenger traffic of 9.7 million.

ACRA's release, as summarized publicly, also says the improvement followed complete repayment of financial debt in 2024, while noting concern about dividend plans in the absence of a formal dividend policy and upcoming major capital expenditures for airport upgrade and expansion. Those two facts belong in the same sentence: the balance sheet may be strong, but airport expansion can re-open capital demands.

For a network-access proposition, the implication is strict opportunity cost. If the airport has to expand capacity, renew equipment, support biometric and passenger-flow projects, manage airfield assets, keep two runways and large terminal systems operating, and possibly fund a next phase toward the long-term concession capacity target, network investment must compete for attention and capital. A switch refresh, redundant edge router, DDoS protection contract or secondary transit circuit may be justified as operational risk management. A speculative paid interconnection program needs a higher bar.

It has to show that incremental ports and service obligations pay for themselves in retained margin or reduce a measurable operating cost elsewhere.

This is where the airport's own digital projects cut both ways. Public releases describe the "Potok" passenger-flow project reducing pre-flight processing time, new operational flow control, electronic boarding-pass gates, and cooperation with IBS on import substitution and domestic ERP solutions. The 2026 anniversary release presents Pulkovo as a technology-forward airport using biometrics and driverless technologies. These projects increase the importance of reliable networks and data systems. They may also increase internal capacity demand. But they are not proof of a telecom resale business.

They may consume capacity and technical staff rather than produce saleable external access revenue.

Customers are concentrated around aviation and airport estate needs

The customer base available to Northern Capital Gateway is attractive but bounded. Airlines are the first obvious group. Pulkovo's year-end release names carrier leaders and says Aeroflot Group companies accounted for large traffic share in rating commentary. The Moscow shuttle program between St Petersburg and Moscow created a high-frequency, high-throughput corridor with more than 70 daily flights announced for the launch and more than 2.2 million passengers carried under the program in 2024. Airline counters, offices, irregular operations desks and ground-handling coordination all need communications.

Yet airlines are sophisticated buyers. A national carrier can procure connectivity through its own telecom contracts, managed WAN providers or mobile/fixed bundles. It will buy airport-managed access where local convenience and controlled-space execution beat direct procurement, not because the airport operator has a unique internet backbone.

Tenants are the second group. Pulkovo's leasing page explicitly seeks commercial partners, discusses business opportunities from daily passenger flows, office rental in the airport business center, advertising and non-aeronautical tariffs. Restaurants, shops, advertisers, offices and service providers need connectivity. Here the airport's access advantage is stronger because the premises are controlled and installation is operationally sensitive. The airport can create value by standardizing onboarding, cabling rules, security approval and support.

But the economics are likely closer to a building or campus managed-service model than a carrier gateway. Revenue is tied to occupancy, tenant churn, commercial-area mix and negotiated lease packages.

Cargo and logistics are a third area, but the boundary is important. The cargo partner page says JSC Pulkovo Cargo Terminal is the largest aviation cargo terminal in Russia's North-West Federal District and provides cargo, mail, terminal processing, consolidation, deconsolidation and customs services. The page does not make Northern Capital Gateway the cargo-terminal operator. EBRD's project description explicitly excludes cargo terminal and business aviation from the airport facilities covered by that concession project.

Cargo-related demand may still appear on the airport estate, but it cannot be counted as NCG network revenue unless a public contract or service offer shows the airport operator charging for that access.

Passengers are the fourth and most tempting group, but passenger connectivity is a dangerous basis for a gateway thesis. Public Wi-Fi and digital passenger services can be important for airport experience and advertising. They can also create costs: access points, controllers, authentication, security monitoring, abuse handling and privacy compliance. Free or sponsored passenger Wi-Fi may support retail dwell time and passenger satisfaction without producing direct network margin. A paid passenger internet model would face user resistance and mobile-data substitutes.

The more realistic passenger-related revenue is indirect: better flow management, digital advertising, parking, premium services and retail conversion. Those benefits may justify internal network investment, but they are not interconnection revenue.

The concentration risk is therefore structural. The airport has many counterparties, but most are tied to one physical estate and one aviation cycle. A shock to travel, security restrictions, airline schedules, sanctions, fuel prices or domestic consumer demand affects the same estate. The network cannot diversify demand the way a citywide ISP can. It can only serve the airport ecosystem unless NCG builds or leases a broader access footprint, for which the public record shows no evidence.

Pricing must separate access control from internet carriage

Northern Capital Gateway already publishes airport pricing surfaces. The partner pages for Russian and foreign airlines provide public airport charge and ground-handling tariff sections, aircraft-class lists, price lists for services at Pulkovo, standard ground-handling agreements and procedures for irregular flights. The cargo motivation page discloses discounts for new or increased cargo flights, including reductions to landing, aviation-security and ground-handling service components. The leasing page points potential tenants toward non-aeronautical service tariffs and commercial-department access.

These are not internet tariffs, but they show a commercial habit: Pulkovo prices the use of controlled airport infrastructure, not only passenger throughput.

A network access product would have to follow the same discipline. The airport can charge for controlled access to airport premises, physical distribution, approved installation, a service demarcation that does not interfere with safety systems, and airport-specific support. It should not pretend that internet transit itself is scarce in St Petersburg. MTS Business, Rostelecom and MegaFon all advertise business connectivity or corporate communications offers in the St Petersburg market. MegaFon is also visible in AS198626's own routing record as AS20632.

ER-Telecom appears as AS9049 in the routing record and as an upstream or peer in third-party network pages. Direct-carrier alternatives cap the price of bandwidth. The airport's defensible price premium must be about execution inside the airport, not raw megabits.

That makes utilisation the decisive number. A port sitting half-empty still needs a switch, router capacity, monitoring, power, cooling, software support and staff knowledge. Even if the airport has already bought upstream bandwidth for its own operations, selling third-party access consumes headroom and creates service promises. If the access product is bespoke, each customer absorbs engineering time. If it is standardized, the operator can push more tenants through the same platform and retain margin from aggregation.

The public record does not disclose port counts, bandwidth commits, tenant take-up, ARPU, churn, support cost or direct telecom revenue. Without those, the honest conclusion is not "profitable" or "unprofitable"; it is "unproven until retained contribution per port is visible."

A practical test would start with five measurements. First, the monthly customer invoice split between external carrier pass-through and NCG-retained service revenue. Second, average and peak utilisation on the access port over a representative period, including event days such as major forums. Third, trouble tickets per customer per month and the share caused inside airport-controlled infrastructure rather than carrier networks. Fourth, equipment and support cost allocated per active customer port over a realistic replacement cycle. Fifth, the direct-carrier quote available to the same customer for equivalent bandwidth and resilience.

If NCG cannot beat direct-carrier procurement after accounting for airport access friction, the product is convenience, not a gateway.

Suppliers and equipment cycles are not neutral

The supplier issue is larger than the names in the BGP table. Routing evidence points to a narrow upstream set. Rating and operating evidence points to upcoming capital expenditure. Pulkovo's digital partnership with IBS is explicitly about import substitution and domestic ERP solutions in the air transport industry. That phrase is a signal of the post-2022 environment: equipment, software, vendor support, compliance and maintenance channels cannot be assumed to follow pre-war European procurement patterns. Fraport's exit and the Russian ownership restructuring add the same warning from the capital side.

The airport still operates, but the supply chain and governance environment is different from the original international-finance concession period.

Network equipment renewal is not a one-time capex line. Routers, switches, optical modules, firewalls, wireless systems, monitoring software, authentication platforms and DDoS protection all age. Vendor support ends. Security updates matter. Spares must be stocked because airport access windows are constrained and failures can become operational incidents. If the operator sells access to third parties, it also inherits customer-facing consequences of these cycles. The airport can defer a nice-to-have feature; it cannot casually defer resilience if the service is promised to airlines, tenants or critical operational partners.

This matters because a gateway business often looks attractive in gross margin before renewal capital. A customer port may ride shared infrastructure with little immediate incremental cash cost. Over a three- to seven-year equipment cycle, however, the operator must replace edge and aggregation gear, add redundancy, maintain staff capability and update security controls. If the customer base is small, the replacement burden falls on few invoices. If the customer base is growing but heterogeneous, support complexity rises.

The winning version is boring and standardized: a few service tiers, clear demarcation, limited custom work, explicit installation charges and priced resilience options. The losing version is prestige integration work sold cheaply because the customer is already a tenant.

The airport's own operating model suggests why management may be tempted to keep network services close. The company already coordinates safety, security, leasing, aviation services, passenger flows and major event surges. During SPIEF 2025, Pulkovo reported about 300,000 passengers over the forum period, 288 flights in four days, strong business-lounge usage growth and more than 400 additional employees engaged. A service outage during such peaks would be costly in reputation and operational stress. That supports internal control. It does not by itself support external expansion.

Competition is direct, indirect and often cheaper

The direct competitors are not mysterious. Corporate customers in St Petersburg can seek office internet, VPN, mobile, fixed and managed communication services from national or regional telecom providers. MTS Business advertises business internet, office internet, VPN, DDoS protection and related services for the region. Rostelecom advertises corporate VPN service in St Petersburg with speeds up to 10 Gbit/s. MegaFon advertises corporate services in St Petersburg including office internet, virtual PBX, IoT and communications services.

These offers do not prove availability at every airport address, but they prove that buyers have alternatives and benchmarks.

The indirect competitor is the customer's own architecture. A retailer can use mobile backup. An airline can extend a managed WAN. A bank terminal can use a payment processor's approved connectivity path. A digital-advertising operator can bring its own connectivity package or negotiate access as part of an advertising concession. A hotel or business-center tenant can ask for direct carrier service if the building access rules allow it. The airport can slow or simplify physical entry, but it cannot repeal the economic comparison.

There is also a substitute in procurement design. NCG can bundle connectivity into leases or operational-service packages, making the telecom line invisible. Bundling can improve take-up because tenants care less about itemized bandwidth. It can also hide whether the network service is profitable. If the tenant pays higher rent and expects working connectivity as part of occupancy, the margin belongs to the whole property package, not to the autonomous system. That may be the right commercial model for an airport, but it should not be counted as proof that a standalone access-port business works.

The best economic position is therefore not to chase the city ISP market. It is to treat the airport as a controlled campus where NCG can aggregate demand that is expensive for outsiders to serve because of security, physical access, service windows and coordination. The operator's moat is not fiber in St Petersburg. It is operational authority at Pulkovo. The product should be priced as airport-grade access and support, not as commodity internet.

Regulation, ownership and sanctions raise the hurdle

Northern Capital Gateway's regulatory perimeter is unusually dense. The Pulkovo disclosure page says the operator provides public access to information under Russian laws and government or regulator instruments covering electric power, natural monopolies, water supply, heat supply and workplace assessment, and links to airport safety, utility, airport-services, electricity, cost-structure and nondiscriminatory-access materials. The public-private partnership page says Pulkovo Airport company monitors NCG's quality of airport services and compliance with development and operation obligations.

The company is therefore used to regulated infrastructure disclosure. That helps discipline. It also means any network-adjacent service must not interfere with higher-order airport obligations.

Ownership has also shifted under pressure. Fraport's own page still describes Northern Capital Gateway's area of activity as operation and expansion of Pulkovo Airport and says the concession became effective on 29 April 2010 for 30 years, while listing a 25 percent Fraport share. Interfax reported in December 2024 that Fraport had completed the sale of its 25 percent stake in VVSS Holding, the parent of Pulkovo's airport operator, to Oman's Orbit Aviation, after a presidential order permitted the sale.

Interfax also described the December 2023 transfer of NCG to VVSS Holding from Cyprus-based Thalita Trading and listed the earlier international shareholder structure. AK&M and FlightGlobal coverage of the 2023 decree emphasized the creation of VVSS Holding and transfer of the airport operator's share capital under special economic measures. Financial media later described Fraport's exit as the end of a 14-year Russian engagement.

Sanctions and counter-sanctions matter even if the network itself is small. OpenSanctions records the company as subject to Ukrainian sanctions, with data connected to the Ukraine NSDC sanctions register and Russian corporate registry enrichment. This does not say the company is sanctioned by every jurisdiction, and it should not be overstated. It does affect counterparties, suppliers, banks and insurers that screen Russian aviation infrastructure exposure. A foreign vendor, cloud provider, airline supplier or equipment distributor may avoid involvement.

A network-service business whose margin depends on imported equipment, foreign support or international counterparties must price that friction.

Regulation can also affect demand. Pulkovo's international traffic grew in selected corridors in 2024, including Turkey, China, Thailand, Uzbekistan, the UAE and Belarus, while western routes remain shaped by geopolitical restrictions. Traffic mix affects the tenants, carriers and service providers physically present at the airport. If airline mix becomes more concentrated around domestic and friendly-country carriers, the accessible customer base may be easier to coordinate but less diverse. If international traffic broadens, service expectations and technical requirements may rise. The network product must survive both scenarios.

Unofficial signals should be read as smoke, not proof

The useful unofficial signals are modest. IPinfo tags at least one IP assigned to AS198626 as VPN and shows no hosted domains on the ASN, no downstreams and one pingable IP in a recent scan. That does not prove a commercial VPN operation. It may reflect remote access, security tooling, misclassification, a passenger or contractor endpoint, or ordinary enterprise use. It is a market signal only because it reminds us that the AS is used for live services, not just a registry ornament.

PeeringDB's absence is another signal. A company can have private transit and no PeeringDB profile. Many enterprise ASNs do. But a paid interconnection seller usually wants to reduce discovery friction. No PeeringDB profile, no visible exchange presence and no downstream cone imply that if paid access exists, it is probably local, private and campus-like. That reinforces the contribution test. A hidden customer base may be real, but it should not be assumed at wholesale scale.

Media reports about financing are also signals. The Moscow Times, citing Bloomberg and people with knowledge of the 2010 deal, reported that Pulkovo financing was expensive relative to emerging-market peers, with a commercial tranche priced hundreds of basis points over Euribor. Old financing spreads do not price today's network project, especially after debt repayment. They do, however, show that the original concession carried country, project and execution risk. A small access business cannot ignore the cost of capital simply because the current balance sheet is stronger.

The official and semi-official records also contain contradictions or stale surfaces. Pulkovo's partner page advertises more than 160 destinations and more than 70 airlines, while the 2024 annual release reports 111 destinations and 55 airlines. Fraport's page still lists a share even though Interfax reports the sale was completed. NIB's 2010 page discusses capacity expectations to 2039 that differ from the Pulkovo PPP page's 35 million passenger target by 2039. These are not fatal contradictions; they reflect different dates and purposes.

For investment judgment, they argue for using the dated operating release and current registry or transaction reports when making hard claims.

The base case is a utility, not a growth story

The base case should be deliberately plain. Northern Capital Gateway needs a strong network because a modern airport cannot operate as a collection of disconnected counters, gates, car parks, security posts, tenant kiosks and management systems. The company has enough scale to justify internal routing control, public address space and serious operational discipline. Its airport estate is large, its passenger flows are meaningful, its event surges are real, and its digital projects make communications more important over time. That is a good utility thesis.

It says the network deserves budget, management attention and renewal capital because the airport cannot run well without it.

It is not yet a growth thesis. Growth requires a repeatable customer proposition outside the airport's own internal requirements. The easiest mistake is to look at the passenger count and imagine a proportional telecom opportunity. The harder truth is that passengers are not stable access customers, airlines are professional buyers with their own networks, tenants can compare direct-carrier quotes, and many commercial systems are bundled into leases or service concessions. The airport can still win business, but only where it solves a coordination problem that a carrier cannot solve cheaply.

That means the service is local, controlled and operationally specific. It is not the same product as generic business broadband.

The capital allocation rule follows from that distinction. Spending to protect airport operations can be approved on resilience grounds even if there is no separate customer invoice. Spending to build a paid access platform needs a unit case. A new redundant edge, a higher upstream commit, a managed Wi-Fi expansion, a tenant aggregation switch or a second carrier handoff should be tied either to measurable operating-risk reduction or to specific customers paying for the incremental burden. If neither condition exists, the network program can become a quiet subsidy from airport economics to a small telecom ambition.

This is also where ownership and sanctions uncertainty meet ordinary engineering economics. A vendor replacement that would once have been routine may now need domestic substitution, extra testing, longer procurement windows or higher spares inventory. A carrier path that looks redundant on paper may share ducts, facilities or operational dependencies in practice. A tenant promise that seems easy during normal weeks may become expensive during a forum, disruption or security event. The airport's strength is command of the premises; its vulnerability is that everything happens on the same premises.

An access business built on that estate must price concentration, not ignore it.

So the most useful near-term management question is not how many services can be sold. It is which services can be refused. A disciplined airport network operator should be able to decline bespoke customer requests that would create disproportionate support burden, refuse low-margin resale where the customer can buy directly from a carrier, and avoid long contract terms that underprice equipment refresh. The service should be standardized enough that the second customer is cheaper to serve than the first. If every new customer needs custom engineering, custom security approval and custom fault escalation, scale is an illusion.

What would change the judgment

The judgment would improve quickly with concrete evidence. A public or verifiable tariff for tenant connectivity, a list of service tiers, a meaningful count of active customer ports, disclosed direct telecom revenue, visible customer references, additional upstreams, IPv6 deployment, an exchange presence, a PeeringDB profile, RPKI clarity and a documented service-level model would show that NCG is managing the airport network as a deliberate access platform. Even stronger would be proof that the retained contribution after carrier pass-through funds equipment renewal and support on a customer-port basis. That is the missing number.

The judgment would worsen if evidence showed that the AS exists only for internal web, VPN or operational addressing; that tenant connectivity is bundled into rent with no separate margin discipline; that direct carriers can reach tenants without airport friction; that the airport is dependent on one upstream for practical traffic despite nominal multi-homing; that network equipment renewal is deferred because airport capex absorbs capital; or that sanctions and supply constraints raise maintenance costs faster than access revenue grows.

It would also worsen if Aeroflot Group or another major counterparty concentration became so dominant that the operator's commercial bargaining power weakened.

The neutral case is the most plausible one. Northern Capital Gateway should be valued as an airport operator with a live autonomous system and campus-network optionality. The AS has operational value. It may support paid access on airport territory. But the public record does not justify calling the company a scalable interconnection business. A gateway is only profitable when it aggregates traffic that customers cannot buy more cheaply and conveniently elsewhere. Pulkovo has the physical control to create that condition in pockets.

It has not publicly shown enough routing diversity, customer depth or pricing evidence to prove that the condition exists at scale.

The managerial answer is not to abandon the network. It is to keep the commercial thesis narrow. Build resilience for the airport first. Standardize tenant access where controlled premises make NCG the efficient coordinator. Charge installation and support explicitly. Treat upstream bandwidth as pass-through unless oversubscription and shared services create measurable contribution. Measure utilisation at the port level. Avoid bespoke connectivity promises that turn a small revenue line into a critical-service liability.

Do not count passenger Wi-Fi, digital transformation or airline traffic as telecom margin unless invoices and costs prove it.

The first paid port is therefore the correct economic instrument. If it is priced as airport-grade access, filled to realistic utilisation, supported with few tickets, protected by route diversity and renewed from retained contribution, it can justify adding the second and third ports. If it merely resells a carrier circuit with airport overhead attached, the business should remain an internal utility. Northern Capital Gateway's advantage is not being a citywide ISP. Its advantage is control of a critical airport estate. The only gateway worth building is one whose customers pay for that control after the carriers have already been paid.

Sources